Production of bio-oil via hydrothermal liquefaction of birch sawdust

Energy Conversion and Management 144 (2017) 243–251

Contents lists available at ScienceDirect

Energy Conversion and Management journal homepage: www.elsevier.com/locate/enconman

Production of bio-oil via hydrothermal liquefaction of birch sawdust Kristaps Malins Institute of Applied Chemistry, Riga Technical University, Paula Valdena Str. 3, LV-1048 Riga, Latvia

a r t i c l e

i n f o

Article history: Received 9 February 2017 Received in revised form 13 April 2017 Accepted 15 April 2017

Keywords: Birch sawdust Lignin Cellulose Hemicellulose Bio-oil Hydrothermal liquefaction

a b s t r a c t The effect of weight ratio of plywood manufacturing by-product birch sawdust (BS) to water (1/2–1/8), reaction temperature (200–340 °C), initial H2 pressure (0–10 MPa), residence time (5–90 min), catalysts amount (0.25–7.0 wt.%) and type (FeSO4, ZnSO4, NiSO4, Raney-nickel, Ni65%/SiO2AAl2O3, Na2CO3 and NaOH) on hydrothermal liquefaction of BS was investigated. High yield of bio-oil (54.1%) with calorific value (CV) 24.9 MJ/kg under developed optimal experimental conditions in the presence of NaOH (5 wt.%) utilizing weight ratio of BS to water 1/4, residence time 5 min, mixing speed 250 rpm at 300 °C without pressurized particular inert gas or H2 atmosphere was achieved. Compounds in bio-oil analyzed by gas chromatography-mass spectrometry (GC-MS) have suitable chemical structures for conversion into renewable hydrocarbons. Marketable solid residue (SR) with yield 7.1%, high CV (29.8 MJ/kg) and perspective characteristics for industrial application was obtained. Produced gas in process analyzed by gas chromatography-thermal conductivity detector (GC–TCD) contains 60.1 vol.% of CO2. Ó 2017 Elsevier Ltd. All rights reserved.

1. Introduction Since the industrial revolution, demand and consumption of fossil sources dramatically increased and caused environmental pollution [1]. Biomass is one of the most widespread renewable energy sources. Bioethanol, biodiesel and hydrotreated vegetable oil already has been produced in large amounts worldwide mainly from food crops for partial substitution of fossil transportation fuels and CO2 emissions reduction [2,3]. Lignocellulosic agriculture and wood industry wastes are available, low cost feedstock for energy production. Furthermore, lignocellulosic biomass has high potential for synthesis of alternative fuels and/or its precursors (synthesis gas and bio-oil) by widely investigated thermochemical conversion processes – gasification, pyrolysis and liquefaction. Lignocellulosic biomass generally contains cellulose, hemicellulose, lignin and extractives. The components in lignocellulosic materials vary according to biomass feedstock type, growth stage and growing conditions of the plant. Extractives are fatty acids, sugars, proteins, fats, phenols, resins, resin acids, terpenes, etc., soluble in organic polar or non-polar solvents. Polysaccharides cellulose and hemicellulose consists from various pentose, hexose and uronic acid monomers, but macromolecule of lignin contains different phenol derivative units [4]. The monomers and polymers in lignocellulosic material are linked by ether, ester, hydrogen and CAC bonds. At high temperature the chemical bonds of macromolecules and other substances in lignocellulosic material can be E-mail address: [email protected] http://dx.doi.org/10.1016/j.enconman.2017.04.053 0196-8904/Ó 2017 Elsevier Ltd. All rights reserved.

broken down to fragments with a wide range of molecular weight distribution [5]. Catalytic hydrothermal liquefaction is one of the most promising thermo-chemical conversion paths of biomass into bio-oil with high calorific value (CV) and chemical structure suitable for liquid renewable aliphatic and/or aromatic hydrocarbon production [6]. The process is similar to geological formation of fossil sources, except bio-oil synthesis from biomass in autoclave-reactor can be performed at low exposure time expressed in hours or even minutes [1]. The hydrothermal experiments typically are carried out at high temperature (up to 400 °C) and pressure (up to 40 MPa) in the aqueous medium [7]. The biooil formation by hydrothermal liquefaction is complex process and degradation mechanisms comprise the following three general steps – depolymerization of the biomass, decomposition of monomers and recombination of reactive fragments. Under hydrothermal conditions at high temperature and pressure the water acts as a solvent and participates as a reactant in hydrolysis reactions of lignocellulosic material [1,8]. Further hydrothermal degradation of fragmented compounds produced by rapid hydrolysis of polysaccharides leads to formation of wide range of different organic compounds, such as, furan derivatives, diols, carboxylic acids, alcohols etc. [9,10]. Depolymerization of lignin by hydrolysis reactions of ether-bonds producing different phenol derivatives is similar to cleavage of carbohydrates [7]. Then depolymerizated fragments of lignocellulosic material are exposed to dehydration, dehydrogenation, decarboxylation and deoxygenation reactions forming compounds with low molecular weight. The benzene ring has a high thermal and chemical stability allowing only further

244

K. Malins / Energy Conversion and Management 144 (2017) 243–251

degradation, elimination and rearrangement of substituent groups at hydrothermal conditions. Reactive fragment rearrangement, condensation, isomerization, cyclization, polymerization, recombination and other chemical transformation leads to new compounds [5]. Hydrothermal liquefaction experiments of biomass in scientific literature usually are performed in inert He [11] and N2 [9,12,13] atmosphere. There are also some reports about bio-oil production without utilization of certain initial gas pressure [14] and abundant number of investigations performed in compressed H2 atmosphere [11,15]. The influence of reducing H2 atmosphere on biomass conversion are extensively investigated in presence of different supported transition (Ni, Mo, Co) and platinum group metal catalysts. Commercial heterogeneous metal catalysts with similar characterizations are widely used in chemical and petroleum refining industry. High pressure reducing atmosphere may provide chemical accession of H2 to fragmented compounds of lignocellulosic material under hydrothermal conditions [7]. An H2 involvement in biomass conversion processes improves the chemical structure of compounds in bio-oil and increases its CV. There are some reports in literature about effect of various water miscible metal Fe(II) Zn(II), Ni(II), Co(II), Mg(II), Cr(III), and Sn(II) sulfates on bio-oil production by hydrothermal conversion of biomass [16,17]. Furthermore, bio-oil with high yield (up to 61.8%) and CV (up to 35.9 MJ/kg) was obtained in presence of strong alkaline catalysts (NaOH, KOH Na2CO3, K2CO3) [14,18–20]. Water soluble alkaline compounds have strong effect on hydrolysis related reactions of lignocellulosic materials producing smaller molecules exposed to thermal decomposition. The alkaline earth metal oxides CaO and MgO also successfully catalyze bio-oil production from biomass [20,21]. However, CaO and MgO usually delivers lower yield of bio-oil in comparison to strong alkali metal hydroxides, due to lower solubility and basicity of Ca(OH)2, Mg(OH)2 formed by hydration of oxides in aqueous medium. Thermo-chemical biomass conversion processes are connected with liquid, gaseous and solid product formation. Yield and characteristics of each separate product depends on many factors – chemical composition of biomass feedstock, experimental conditions, including reaction atmosphere, reaction mixture medium and catalyst type. Present study demonstrates the catalysts screening for the biooil production from plywood manufacturing by-product birch sawdust (BS) by hydrothermal liquefaction. Gaseous products (GP), marketable solid residue (SR) and high yield of bio-oil were obtained under developed optimal hydrothermal conditions in presence of best catalyst. Gas chromatography-mass spectrometry (GC-MS), gas chromatography-thermal conductivity detector (GC– TCD), thermogravimetric (TG) and C,H,N,S elemental analysis were utilized for product characterization.

2. Materials and methods 2.1. Materials BS (without bark fraction) was supplied from the local plywood producer Latvijas finieris AS (Latvia) with water content 5.8 wt.%. 6.0 kg of BS was grinded and sieved until particle size
120 mesh. Afterwards the BS was dried at 110 °C for 6 h. The main characteristics of dried and sieved BS sample are given in Table 1. FeSO47H2O, ZnSO47H2O, NiSO47H2O, Na2CO3, NaOH, tetrahydrofuran (THF) with purity of 98 wt.%, concentrated mineral acids HCl (35.7 wt.%), H2SO4 (96.3 wt.%), Raney-nickel (activated nickel catalyst slurry in water, nickel content 89 wt.%), Ni65%/ SiO2AAl2O3 (specific surface area 175 m2/g) and activated carbon (mesh 20–40), CH4 (assay 99.999 wt.%) and CO (assay 99.0 wt. %) were purchased from Sigma–Aldrich Chemie GmbH. The concentrations of mineral acids were determined using volumetric acid-base titration. Acetone (assay 99.7 wt.%) was supplied from Latvijas kimija SIA and H2, CO2 (assay 99.999 wt.%) from Elme messer L SIA.

2.2. Hydrothermal liquefaction of birch sawdust (BS) The effect of various catalysts (FeSO4, ZnSO4, NiSO4, Raneynickel, Ni65%/SiO2AAl2O3, Na2CO3, NaOH), reaction temperature (200, 250, 300, 340 °C), catalyst amount (0.5, 2.5, 5.0, 7.0, wt.% calculated from anhydrous compound of the dried BS amount), weight ratio of BS to water (1/2, 1/4, 1/6, 1/8), initial H2 pressure (2.0, 5.0, 8.0, 10.0 MPa) and residence time (5, 20, 40, 60, 90 min) on hydrothermal liquefaction of BS were investigated. Each experiment was conducted using accurately mixed 40 g of dried BS, catalyst and distilled water in a batch type stainless steel autoclavereactor (Parker Autoclave Engineers Inc.) designed to maximum pressure 37.9 MPa at 343 °C, volume 500 ml, equipped with magnetically coupled mechanical mixer (6 blade agitator impeller) and Sentinel Series Controller. After the filling and sealing the autoclave-reactor was purged with H2 (flow rate 10 ml/s) for 3 min to fully eliminate air atmosphere from the system, when H2 atmosphere was utilized in experiments. Then initial H2 pressure was increased to necessary value. Heating rate was chosen from 3 to 6 °C/min to reach predetermined temperature. From this point of reference the residence time of the hydrothermal liquefaction process of BS was measured. The experiments without initial H2 pressure also were performed to compare the results. Mixing speed of 250 rpm in all the experiments was utilized. After hydrothermal liquefaction process the autoclave-reactor was cooled down to 25 °C with air fan

Table 1 The main characteristics of birch sawdust (BS). Proximate analysis, wt.% (d.b.a) b

b,c

VM

FC

85.6

14.1

Ultimate analysis, wt.% (d.a.fd) Ash

C

H

N

0.3

48.5

6.3


0.3e

S

O

f

45.2

Calorific value (d.b.a), MJ/kg 19.4

Composition analysis, wt.% (d.b.a)

a b c d e f g

Cellulose

Hemicelluloseg

Lignin

Extractives

45.3

24.2

22.9

7.6

On a dry basis. VM - volatile matter, FC - fixed carbon. By difference (FC% = 100%  VM%  Ash%). On a dry and ash free basis. Method detection limit. By difference (O% = 100%  C%  H%). By difference (Hemicellulose% = 100%  Cellulose%  Lignin%  Extractives%  Ash%).

K. Malins / Energy Conversion and Management 144 (2017) 243–251

for 30–50 min. Then autoclave – reactor was depressurized. The volume of the gaseous products (GS) was measured using graduated cylinder with the bottom end immersed in water. The biooil from autoclave reactor was extracted using acetone (500 ml) as solvent. The mixture containing bio-oil, water, SR and acetone was filtered by Büchner funnel. SR was washed with additional 200 ml of acetone and dried at 110 °C for 6 h. The acetone and water from bio-oil were separated by vacuum distillation (0.9– 1.5 kPa) with rotary evaporator at 40–95 °C for 40 min. Bio-oil sample obtained under optimal conditions using NaOH as catalyst has a high ash content. Increased ash content is connected to high amount of Na2CO3, different alkali metal salts (mostly Na) of carboxylic acids and phenol derivatives remained in bio-oil after its extraction from reaction mixture. Bio-oil was treated with concentrated HCl in water/acetone (volume ratio 1/19) solution (pH
1) to reduce ash content. Inorganic salts were removed by centrifugation. The recovery of acetone from sample was carried out in similar manner as in bio-oil extraction procedure from reaction mixture after hydrothermal liquefaction of BS. 2.3. Analysis of feedstock and products from the hydrothermal liquefaction of birch sawdust (BS) Moisture content of BS and SR was determined (drying temperature 105 °C) by EM 120 – HR moisture analyzer (Precisa gravimetrics GA). TG analysis of commercial activated carbon, BS and SR was carried out by Simultaneous Thermal Analyzer (STA) 6000 (PerkinElmer Inc.). Analysis was conducted using N2 flow rate of 10 ml/ min and heating rate of 10 °C/min at maximum temperature 850 °C. Ash content of BS, SR and bio-oil was determined according to ASTM E1755–01 (2007) standard method. C, H, N, S elemental analysis of BS, bio-oil and SR was determined by EA3000 (EuroVector SpA) elemental analyzer. CV (higher heating value) of BS, SR and bio-oil was determined using C 200 (IKAÒ-Werke GmbH & Co. KG) oxygen-bomb calorimeter according to the standard method DIN 51900-3:2005. The extractives content was determined by extraction of BS with 400 ml of acetone. 20 g of the dried BS was refluxed in 150 ml Soxhlet extractor for 10 h. Extractive-free lignocellulosic fraction was dried for 6 h at 110 °C. The extractives content was calculated by difference between initial BS and extractive-free lignocellulosic fraction mass. Lignin content of BS was determined according to modified Klason method (determination of acid-insoluble lignin) [22,23]. 3 g of extractive-free BS sample was mixed with 45 ml of 72% H2SO4 and intensively stirred for 4 h at 25 °C in 1000 ml Erlenmeyer flask. Afterwards the mixture was diluted with 600 ml of distilled water and then refluxed for 4 h. After refluxing the sediment (lignin) was filtered, washed several times with distilled water to neutral pH and dried for 6 h at 110 °C. Cellulose content of BS was determined according to reported analytical procedure [23]. 3 g of the extractive-free BS sample was added to 150 ml of aqueous NaOH solution (17.5%) and intensively stirred at 25 °C for 30 min in 300 ml beaker. The cellulose fraction was separated by filtration, sequentially washed with aqueous NaOH (9.5%), distilled water, aqueous acetic acid (10%) and then again with distilled water to neutral pH. Extracted cellulose sample was dried at 110 °C for 6 h. Hemicellulose content was calculated by the difference (Hemicellulose% = 100%  Cellulose%  Lignin%  Extractives%  Ash%). Chemical composition of bio-oil was identified by GC-MS. Prior to the analysis bio-oil sample was dissolved in THF. GC-MS analyses were conducted by a gas chromatograph GCMS-QP2010 Ultra

245

(Shimadzu Corp.) equipped with RTX-XLB capillary column (60 m  0.25 mm  0.25 mm), a quadrupole mass selective analyzer and secondary electron multiplier detector. The flow rate of carrier gas (He) was 1.0 ml/min. The temperature of injector and of ion source was 320 and 280 °C, respectively. Temperature program: 50 °C ? 200 °C (7 °C/min, hold 10 min) ? 320 °C (5 °C/min, hold 10 min). Total program time was 65.4 min. The compounds were identified by comparing their mass spectra (35–500 m/z) with the NISTO8 database. Calibration was not carried out, due to the large number of compounds presented in the analyzed samples. A comparative evaluation of the samples was performed by integrating the total ion current (TIC) chromatogram and calculating the area percentage of each identified peak. These data are not absolute and serve only for purposes of comparison and evaluation. Part of components in GS obtained under optimal hydrothermal liquefaction conditions of BS were determined by GC–TCD. Analyses were carried out utilizing a gas chromatograph GC-2010 Plus (Shimadzu Corp.) equipped with a Carboxen 1010 PLOT capillary column (30 m  0.32 mm  0.15 mm). The flow rate of carrier gas (N2) in column was 20 ml/min. Gaseous sample was taken from autoclave-reactor after hydrothermal liquefaction experiment. The temperature of injector and detector were 200 and 225 °C, respectively. Total program time at column temperature 100 °C was 15 min. Peaks of gaseous components were identified by comparison to reference gases CO2, H2, CH4 (assay 99.999 wt.%) and CO (assay 99.0 wt.%). CO2 and H2 content by volume of each identified peak was determined using five different calibration amounts (5–200 ml) of reference gases at constant temperature 23 °C. The coefficient of determination (R2) for CO2 and H2 calibration lines were 0.9945 and 0.9968. Maximum relative standard deviation (RSD) of analytical results was
5.0%. Special purpose 250R-V-GT (SGE Analytical Science) 5–250 ml pre-fit syringe with removable needle push-pull valve was used for injection of samples and reference gases. CO and CH4 content by volume were not determined, due to the lower amount in GP than in minimum calibration sample 5 ml (2%).

2.4. Calculations Energy recovery ER calculated by Eq. (1) is an important parameter to evaluate the influence of experimental conditions on overall transformation of BS into desired products – bio-oil and/or energetically valuable char containing solid residue [24].

ER ð%Þ ¼ CVP  YP =CVF

ð1Þ

where CVP and CVF are calorific values of product and feedstock, respectively. YP (%) is yield of product and calculated by Eq. (2).

Yp ð%Þ ¼ mP =mF  100

ð2Þ

where mP is the mass of product and mF is the mass of feedstock. Each hydrothermal liquefaction experiment of BS was repeated twice. The yield of bio-oil, solid product and produced amount of GS was determined as arithmetic average from the two independently repeated experiments. A third independent experiment was performed in cases when previous two experimental values differed more than 6%. The average result of experiment with maximum RSD 4.0% was determined using two closest experimental values. The absolute values of all quantitative analysis data were obtained in similar manner and the maximum RSD was 3.0%. The maximum RSD of GP analysis determined by GC-TCD was 5.0%.

246

K. Malins / Energy Conversion and Management 144 (2017) 243–251

3. Results and discussion 3.1. Effect of various catalysts and reaction conditions on the hydrothermal liquefaction of birch sawdust (BS) Fig. 1 shows the effect of various catalysts on hydrothermal liquefaction of BS. The lowest energy recovery (21.2%) and yield of bio-oil (22.3%) was obtained without catalyst. In the presence of FeSO4 and ZnSO4 energy recovery slightly increased to 24.8 and 27.2%, respectively. NiSO4 from all weakly acidic salts showed the highest effect on BS conversion to bio-oil, with energy recovery of 45.5%. Significantly higher energy recovery (63.1–65.8%), yield of biooil (44.0–44.2%) and CV (27.8–28.8 MJ/kg) was achieved using high dispersion Raney-nickel powder and Ni65%/SiO2AAl2O3. This observation suggests that nickel based materials containing particles of elemental metal are perspective catalysts for bio-oil production from lignocellulosic and other type biomass feedstock by hydrothermal liquefaction providing high yield of bio-oil [14,25,26]. However, the highest energy recovery (68.4–68.8%) and yield of bio-oil (48.6–53.9%) was achieved using alkaline catalysts (Na2CO3 and NaOH). The yield of bio-oil and CV (24.7– 27.3 MJ/kg) is close for both alkaline catalysts and depends on water and ash content in bio-oil. It can be changed by additional drying of bio-oil or reduction of ash content. Alkaline catalysts have strong impact on hydrolysis and other depolymerisation related reactions providing high yield of bio-oil [27]. Different compounds with acidic properties in feedstock as well as produced in hydrothermal liquefaction process react with alkaline catalysts forming different organic alkali metal salts [7,28]. Alkali metal salts mostly carboxylic acid and phenol derivatives produced from biomass usually are very susceptible to thermo-chemical transformation at high temperature and easily decomposes to less polar organic chemical compounds with lower oxygen content than feedstock or its hydrolysis products. Obtained results indicate that strong homogeneous water soluble alkaline catalysts can be successfully used in hydrothermal processing of lignocellulosic material. Utilization of NaOH for hydrothermal liquefaction has advantage in comparison to Na2CO3. NaOH is significantly more basic than Na2CO3 and it reacts with part of CO2 formed in hydrothermal process producing Na2CO3, which also acts as catalyst. Thermo-chemical conversion processes of lignocellulosic or other kinds of biomass are connected with formation of high CO2 amount [7,29] in the decarboxylation reactions.

The effect of reaction temperature on hydrothermal liquefaction of BS is illustrated in Fig. 2. The lowest energy recovery (39.4%) and yield of bio-oil (49.9%) was acquired at 200 °C. This can be explained by reduced rate of hydrolysis and thermal decomposition connected reactions of organic matter in BS at lower temperature. Incomplete thermal degradation of compounds with high oxygen content produced in hydrolysis reactions of BS leads to the bio-oil with low CV (15.3 MJ/kg). The increase of reaction temperature significantly improved bio-oil production process. The highest energy recovery (68.8%) and yield of bio-oil (53.9%) was obtained at 300 °C. Further increase of reaction temperature has adverse effect on bio-oil formation, due to more pronounced thermal decomposition of feedstock and bio-oil to elevated amount of GP and SR. Similar observations were described in the research paper [18], when alkaline catalysts (NaOH, Na2CO3, K2CO3) for subcritical water liquefaction of lignocellulosic feedstock was used. The energy recovery decreased to 39.9% and yield of bio-oil to 23.2% at 340 °C. Moreover, under these experimental conditions the bio-oil with highest CV (33.3 MJ/kg) was obtained. This indicates the presence of substances with low oxygen content in bio-oil formed by excessive thermal degradation. Increase of residence time reduces energy recovery and yield of bio-oil (Fig. 3). Furthermore, CV of bio-oil (24.7–25.6 MJ/kg) obtained in 40–90 min was slightly higher in comparison to CV of bio-oil (23.3–23.8 MJ/kg) extracted at shorter residence time (5–20 min). This implies that excessive thermal degradation of feedstock similarly to the effect of high reaction temperature occurs at long residence time. Relatively large part of BS decomposes to GP and SR decreasing the energy recovery and yield of bio-oil. Under studied experimental conditions the highest energy recovery (74.7–75.1%) and yield of bio-oil (61.2–62.0%) was achieved at residence time from 5 to 20 min. The effect of weight ratio of BS to water is depicted in Fig. 4. The highest energy recovery (73.8–74.2%) and yield of bio-oil (56.7– 58.1%) was reached utilizing weight ratio of BS to water from 1/6 to 1/8. The water amount has a high impact on biomass defragmentation reactions and may involve in chemical accession to depolymerized biomass molecules increasing yield of bio-oil and energy recovery. Biomass to water ratio also has significant effect on product distribution and composition [30]. On the other hand, excessive water amount considerably complicates the bio-oil extraction and drying [31]. Increased amount of polar solvent is necessary to complete dissolution of bio-oil before its separation

Fig. 1. The effect of various catalysts on bio-oil production from birch sawdust (BS) by hydrothermal liquefaction. Experimental conditions: weight ratio of BS to water 1/4, reaction temperature 300 °C, initial H2 pressure 5.0 MPa, residence time 40 min, catalyst amount 5 wt.% of dry BS, mixing speed 250 rpm.

K. Malins / Energy Conversion and Management 144 (2017) 243–251

247

Fig. 2. The effect of reaction temperature on bio-oil production from birch sawdust (BS) by hydrothermal liquefaction. Experimental conditions: weight ratio of BS to water 1/ 4, initial H2 pressure 5.0 MPa, residence time 40 min, catalyst (NaOH) amount 5 wt.% of dry BS, mixing speed 250 rpm.

Fig. 3. The effect of residence time on bio-oil production from birch sawdust (BS) by hydrothermal liquefaction. Experimental conditions: weight ratio of BS to water 1/4, reaction temperature 300 °C, initial H2 pressure 5.0 MPa, catalyst (NaOH) amount 5 wt.% of dry BS, mixing speed 250 rpm.

Fig. 4. The effect of weight ratio of birch sawdust (BS) to water on bio-oil production from BS by hydrothermal liquefaction. Experimental conditions: reaction temperature 300 °C, initial H2 pressure 5.0 MPa, residence time 40 min, catalyst (NaOH) amount 5 wt.% of dry BS, mixing speed 250 rpm.

from SR. Dilution of polar solvent with high amount of water reduces dissolving ability of bio-oil. This leads to increased energy consumption caused by high amount of water, polar solvent evaporation and recovery in bio-oil extraction and drying step, especially when weight ratio of BS to water 1/6 was used. Thereby,

recommended weight ratio of BS to water is 1/4 to facilitate biooil extraction procedure and reduce overall energy consumption. Likewise to the weight ratio of BS to water catalyst amount has a similar effect on hydrothermal liquefaction process (Fig. 5). This effect can be explained by intense involvement of NaOH in hydrol-

248

K. Malins / Energy Conversion and Management 144 (2017) 243–251

Fig. 5. The effect of catalyst (NaOH) amount on bio-oil production from birch sawdust (BS) by hydrothermal liquefaction. Experimental conditions: weight ratio of BS to water 1/4, reaction temperature 300 °C, initial H2 pressure 5.0 MPa, residence time 40 min, mixing speed 250 rpm.

ysis and other defragmentation related reactions of BS in water environment, where alkali also acts as reactant, forming different organic and inorganic salts. The optimal energy recovery and yield of bio-oil was obtained when catalyst amount 5.0 wt.% was utilized. Further increase of catalyst amount to 7.0 wt.% gave insignificant influence on overall conversion of BS to bio-oil and difference between both energy recoveries was only 0.3%. NaOH amount used in process affects bio-oil ash content and CV. The bio-oil with the lowest CV (23.8 MJ/kg) was obtained utilizing highest catalyst amount. The effect of initial H2 pressure on bio-oil production from BS is showed in Fig. 6. The moderate increase of energy recovery and bio-oil yield was observed at higher initial H2 pressure. Energy recovery (66.8%) obtained without utilization of compressed H2 atmosphere was 3.5% lower than energy recovery achieved at highest initial H2 pressure (10.0 MPa). This effect can be explained with increased amount of liquid phase towards vapor (gaseous phase) under high pressure providing more robust chemical interaction of water and catalyst with solid lignocellulosic feedstock and condensed liquid products obtained during the process. In the result increase of energy recovery and yield of bio-oil was observed at elevated pressure. The energy recovery up to 75% can be obtained at initial pressure H2 of 5.0 MPa and shorter residence

time than
20 min (Fig. 3). These observations indicate that initial H2 pressure has significant impact on bio-oil production process, but to reach maximum energy recovery the residence time should be in range from 5 to 20 min. However, specially designed expensive high pressure equipment and robust safety requirements are necessary for storage and operation with pressurized gases, particularly with H2 in chemical production. Therefore, recommended optimal hydrothermal liquefaction conditions of BS conversion into valuable products are weight ratio of BS to water 1/4, reaction temperature 300 °C, residence time 5 min, catalyst amount 5 wt.% of dry BS, mixing speed 250 rpm, without utilization of initial H2 pressure. Under these experimental conditions BS was converted to bio-oil with high yield (54.1%) and energy recovery (69.6%) without utilization of highly compressed specific inert gas or H2 atmosphere. Vapors of hot compressed water, GP and liquid organic compounds with low molecular weight and boiling point produced by thermo-chemical decomposition of feedstock raised the pressure in autoclave-reactor up to 8.5–9.1 MPa at the end of the process. It dropped to 0.90 MPa, when autoclave-reactor was cooled down to room temperature. The volume of produced GP was 3.6 L (measured at 25 °C), but yield of SR was 7.1%. Overall energy recovery from BS to liquid and solid products reached 80.5%. The rest energy in organic matter in form of feed-

Fig. 6. The effect of initial H2 pressure on bio-oil production from birch sawdust (BS) by hydrothermal liquefaction. Experimental conditions: weight ratio of BS to water 1/4, reaction temperature 300 °C, residence time 40 min, catalyst (NaOH) amount 5 wt.% of dry BS, mixing speed 250 rpm.

249

K. Malins / Energy Conversion and Management 144 (2017) 243–251 Table 2 The main characteristics of bio-oil samples and solid residue (SR) obtained under optimal hydrothermal liquefaction conditions from birch sawdust (BS). Ultimate analysis, wt.% (d.a.fb)

Bio-oil SR Treated bio-oile a b c d e

C

H

N

64.5 74.6 64.4

6.2 5.3 6.6


0.3c

S

O

Ash (d.b.a), wt.%

Calorific value (d.b.a), MJ/kg

6.8 1.6 0.5

24.9 29.8 27.1

d

29.3 20.1 29.0

On a dry basis. On a dry and ash free basis. Method detection limit. By difference (O% = 100%  C%  H%). Bio-oil after treatment with concentrated HCl in water/acetone (volume ratio 1/19) solution, separation of inorganic salts by centrifugation and solvent recovery.

stock was transferred to different gases and compounds with low boiling point and lost during the bio-oil extraction process from reaction mixture. 3.2. Characterization of products obtained by hydrothermal liquefaction of birch sawdust (BS) The main characteristics of liquid and solid hydrothermal liquefaction products obtained under optimal conditions from BS are given in Table 2. The bio-oil is highly viscous dark brown color liquid substance, well soluble in polar solvents such as acetone, THF and 1,4-dioxane. H content of feedstock (Table 1) and bio-oil sample was relatively close, but C content increased by 16% in comparison to BS. The bio-oil has relatively high ash content which can be reduced by simple treatment method with concentrated HCl in water/acetone (volume ratio 1/19) solution. NaCl formed in HCl reaction with alkaline catalyst and sodium salts from reaction mixture has low solubility in acetone and can be easily separated by centrifugation. The ash content in treated bio-oil after acetone recovery was decreased by 93%, but CV increased by 2.2 MJ/kg. Organic matter loss caused by deashing procedure reduced overall energy recovery of treated bio-oil to 61.4%. The bio-oil contains large amount of different phenol derivatives mostly based on pyrogallol chemical structure produced from lignin and approximately two times less cellulose, hemicellulose derived compounds (Table 3). The bio-oil also contains low amount of by-products produced in alkaline catalyzed self-aldol condensation reactions of acetone. Furthermore, 0.8% of 2,2,4-trimethyl1,3-dioxolane was identified in treated bio-oil sample. 1,3Dioxolane derivatives produces in acid catalyzed reactions of acetone with 1,2-diols during bio-oil treatment process in presence of the concentrated HCl. Other evident differences in chemical composition of both bio-oil samples were not observed. The acetone derived by-product amount can be reduced to minimum by proper vacuum drying of bio-oil if necessary. THF or 1,4-dioxane can be successfully utilized instead of acetone for bio-oil extraction from reaction mixture, due to high dissolution ability and miscibility in water. These solvents have significantly lower chemical interaction ability with chemical compounds in bio-oil than acetone. However, acetone is inexpensive and available solvent with low boiling point. The compounds in obtained bio-oil have appropriate chemical structure for aliphatic and aromatic hydrocarbon synthesis by catalytic hydrotreatment and hydrocracking [32–34]. Bio-oil upgrading to hydrocarbons or substances with low oxygen content can be performed over Ni based catalysts [35,36]. Ni based catalysts improves physicochemical properties of bio-oil in subcritical, supercritical water [37,38] and the presence of alcohol at high temperature [39]. SR is the second general hydrothermal liquefaction product. It is black color solid substance with high CV and relatively low ash content. TG analysis curves of BS, SR and commercial activated carbon as a reference are illustrated in Fig. 7.

Table 3 The main compounds of bio-oil obtained under optimal hydrothermal liquefaction conditions from birch sawdust (BS) and identified by gas chromatography-mass spectrometry (GC-MS). No. Compound Cellulose and hemicellulose derived products 1 Acetic acid 2 2-Propanol, 1-methoxy3 n-Propyl acetate 4 Propanoic acid 5 Acetic acid, hydroxy-, methyl ester 6 1,2-Propanediol 7 1,2-Butanediol 8 2-Furanmethanol, tetrahydro9 2-Cyclopenten-1-one, 2-methyl10 Butyrolactone 11 2(3H)-Furanone, dihydro-5-methyl12 2-Cyclopenten-1-one, 3-methyl13 2-Cyclopenten-1-one, 2,3-dimethylSub-total Lignin derived products 15 Phenol, 2-methyl16 Phenol, 2-methoxy17 Phenol, 2,3,6-trimethyl18 Phenol, 2-methoxy-4-methyl19 Phenol, 4-ethyl-2-methoxy20 Phenol, 2,6-dimethoxy21 1,2,4-Trimethoxybenzene 22 Benzene, 1,2,3-trimethoxy-5-methyl23 2-Propanone, 1-(4-hydroxy-3methoxyphenyl)24 Phenol, 2,6-dimethoxy-4-(2propenyl)Sub-total Acetone derived by-products 25 3-Penten-2-one, 4-methyl26 2-Pentanone, 4-hydroxy-4-methylSub-total

Relative amount in samplea, %

RTb, min

2.5 2.3 1.8 3.0 1.6 3.7 2.3 3.1 2.1 5.6 0.9 1.4 2.9 33.2

10.0 10.4 10.7 11.3 11.9 12.1 14.4 15.3 15.9 16.7 17.7 18.0 18.4

1.4 8.1 0.9 1.3 3.9 37.8 4.7 4.3 1.4

20.1 21.0 23.7 23.9 26.2 29.0 31.8 34.3 35.4

0.8

40.7

64.6 1.3 0.9 2.2

12.8 14.0

a Calculated by the area normalization method, excluding compounds with peak area less than 0.8%. b Retention time.

The maximum weight loss of commercial activated carbon is insignificant in comparison to BS, due to it is almost completely carbonized and has high C content. The form of TG curve and maximum weight loss of SR inclines to commercial activated carbon. Obtained observations suggests that SR is partially carbonized lignocellulosic material with high fraction of carbon. SR with current characteristics can be used as solid fuel and has high potential for industrial application. The large amount of GP was formed in hydrothermal liquefaction process of BS. Part of gaseous component contents by volume in GP was determined by GC-TCD (Fig. S1 in the supplementary data). The dominant component in GP was CO2 (60.1 vol.%) produced in decarboxylation reactions during the hydrothermal lique-

250

K. Malins / Energy Conversion and Management 144 (2017) 243–251

Fig. 7. Thermogravimetric (TG) analysis curves of commercial activated carbon and birch sawdust (BS), solid residue (SR) produced under optimal hydrothermal liquefaction conditions from BS.

faction process of BS. The sample also contains 4.6 vol.% of H2 and traces of CO, CH4, but the rest could be gases with higher molecular weight than light hydrocarbons (C2AC3) whose detection and quantitative analysis are limited using current GC column. Developed optimal hydrothermal liquefaction conditions of BS into energetically valuable materials at relatively low working pressure (up to 9 MPa) in the presence of inexpensive and widely available catalyst NaOH without utilization of highly pressurized particular inert gas or H2 atmosphere have potential for industrial application. There is high possibility that developed optimal conditions could be utilized for various type of lignocellulosic biomass feedstock. 4. Conclusions Alkaline catalysts Na2CO3 and NaOH showed the highest effect on hydrothermal liquefaction of BS in comparison to FeSO4, ZnSO4, NiSO4, Raney-nickel and Ni65%/SiO2AAl2O3. High yield of bio-oil (54.1%) was achieved under developed optimal hydrothermal liquefaction conditions in the presence of NaOH at relatively low operating pressure (up to 9 MPa) without utilization of specific inert gas or reducing H2 atmosphere. The bio-oil has high content of various valuable cellulose, hemicellulose and lignin derived compounds (mostly pyrogallol based different phenol derivatives) with appropriate chemical structure for aliphatic and aromatic hydrocarbon synthesis by catalytic hydrotreatment and hydrocracking. SR (yield 7.1%) is perspective carbonaceous material for utilization as solid fuel and has potential for other industrial application, due to high CV (29.8 MJ/kg) and C content (74.6 wt.%). 90 L (measured at 25 °C) of GP was produced from 1 kg of BS in hydrothermal liquefaction process, where 60.1 vol.% was CO2. This indicates to the dominant decarboxylation reactions. Overall energy recovery of BS to bio-oil and SR with valuable characteristics reached 80.5%. The bio-oil characteristics can be improved by simple treatment method with concentrated HCl in water/acetone (volume ratio 1/19) solution. CV of bio-oil increased from 24.9 to 27.1 MJ/kg and ash content decreased from 6.8 to 0.5 wt.% after separation of inorganic salts by centrifugation and solvent recovery. Overall energy recovery decreased 8% during the treatment procedure. Acknowledgments This work was supported by the National Research Program of Latvia ‘‘LATENERGI”.

Appendix A. Supplementary material Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.enconman.2017. 04.053.

References [1] Tekin K, Karagöz S, Bektasß S. A review of hydrothermal biomass processing. Renew Sustain Energy Rev 2014;40:673–87. http://dx.doi.org/10.1016/j. rser.2014.07.216. [2] Demirbas A. Competitive liquid biofuels from biomass. Appl Energy 2011;88:17–28. http://dx.doi.org/10.1016/j.apenergy.2010.07.016. [3] Boonrod B, Prapainainar C, Narataruksa P, Kantama A, Saibautrong W, Sudsakorn K, et al. Evaluating the environmental impacts of biohydrogenated diesel production from palm oil and fatty acid methyl ester through life cycle assessment. J Clean Prod 2017;142:1210–21. http://dx.doi. org/10.1016/j.jclepro.2016.07.128. [4] Le Floch A, Jourdes M, Teissedre PL. Polysaccharides and lignin from oak wood used in cooperage: composition, interest, assays: a review. Carbohyd Res 2015;417(94–102):6. http://dx.doi.org/10.1016/j.carres.2015.07.003. [5] Balat M. Mechanisms of thermochemical biomass conversion processes. Part 3: Reactions of liquefaction. Energy Source Part A 2008;30(7):649–59. http://dx. doi.org/10.1080/10407780600817592. [6] Dimitriadis A, Bezergianni S. Hydrothermal liquefaction of various biomass and waste feedstocks for biocrude production: a state of the art review. Renew Sustain Energy Rev 2017;68:113–25. http://dx.doi.org/10.1016/j. rser.2016.09.120. [7] Toor SS, Rosendahl L, Rudolf A. Hydrothermal liquefaction of biomass: a review of subcritical water technologies. Energy 2011;36:2328–42. http://dx.doi.org/ 10.1016/j.energy.2011.03.013. [8] Guo Y, Wang SZ, Xu DH, Gong YM, Ma HH, Tang XY. Review of catalytic supercritical water gasification for hydrogen production from biomass. Renew Sustain Energy Rev 2010;14:334–43. http://dx.doi.org/10.1016/j. rser.2009.08.012. [9] Jindal MK, Jhab MK. Effect of process parameters on hydrothermal liquefaction of waste furniture sawdust for bio-oil production. RSC Adv 2016;6:41772–80. http://dx.doi.org/10.1039/c6ra02868c. [10] Herrera R, da Silva DT, Llano-Ponte R, Labidi J. Characterization of pine wood liquid and solid residues generated during industrial hydrothermal treatment. Biomass Bioenerg 2016;95:174–81. http://dx.doi.org/10.1016/j. biombioe.2016.10.006. [11] Duan P, Savage PE. Hydrothermal liquefaction of a microalgae with heterogeneous catalysts. Ind Eng Chem Res 2011;50:52–61. http://dx.doi. org/10.1021/ie100758s. [12] Yang L, Nazari L, Yuan Z, Corscadden K, Xu C, He Q. Hydrothermal liquefaction of spent coffee grounds in water medium for bio-oil production. Biomass Bioenerg 2016;86:191–8. http://dx.doi.org/10.1016/j.biombioe.2016.02.005. [13] Yu G, Zhang Y, Guo B, Funk T, Schideman L. Nutrient flows and quality of biocrude oil produced via catalytic hydrothermal liquefaction of low-lipid microalgae. Bioenerg Res 2014;7:1317–28. http://dx.doi.org/10.1007/s12155014-9471-3. [14] Bi Z, Zhang J, Peterson E, Zhu Z, Xia C, Liang Y, et al. Biocrude from pretreated sorghum bagasse through catalytic hydrothermal liquefaction. Fuel 2017;188:112–20. http://dx.doi.org/10.1016/j.fuel.2016.10.039.

K. Malins / Energy Conversion and Management 144 (2017) 243–251 [15] Li N, Wei L, Bibi R, Chen L, Liu J, Zhang L, et al. Catalytic hydrogenation of alkali lignin into bio-oil using flower-like hierarchical MoS2-based composite catalysts. Fuel 2016;185:532–40. http://dx.doi.org/10.1016/j.fuel.2016.08.001. [16] Raikova S, Smith-Baedorf H, Bransgrove R, Barlow O, Santomauro F, Wagner JL, et al. Assessing hydrothermal liquefaction for the production of bio-oil and enhanced metal recovery from microalgae cultivated on acid mine drainage. Fuel Process Technol 2016;142:219–27. http://dx.doi.org/10.1016/ j.fuproc.2015.10.017. [17] Kong L, Li G, Wang H, He W, Ling F. Hydrothermal catalytic conversion of biomass for lactic acid production. J Chem Technol Biotechnol 2008;83:383–8. http://dx.doi.org/10.1002/jctb.1797. [18] Mazaheri H, Lee KT, Mohamed AR. Influence of temperature on liquid products yield of oil palm shell via subcritical water liquefaction in the presence of alkali catalyst. Fuel Process Technol 2013;110:197–205. http://dx.doi.org/ 10.1016/j.fuproc.2012.12.015. [19] Deniel M, Haarlemmer G, Roubaud A, Weiss-Hortala E, Fages J. Optimisation of bio-oil production by hydrothermal liquefaction of agro-industrial residues: blackcurrant pomace (Ribes nigrum L.) as an example. Biomass Bioenerg 2016;95:273–85. http://dx.doi.org/10.1016/j.biombioe.2016.10.012. [20] Nazari L, Yuan Z, Souzanchi S, Ray MB, Xu C. Hydrothermal liquefaction of woody biomass in hot-compressed water: catalyst screening and comprehensive characterization of bio-crude oils. Fuel 2015;162:74–83. http://dx.doi.org/10.1016/j.fuel.2015.08.055. [21] Yim SC, Quitain AT, Yusup S, Sasaki M, Uemura Y, Kida T. Metal oxidecatalyzed hydrothermal liquefaction of Malaysian oil palm biomass to bio-oil under supercritical condition. J Supercrit Fluids 2017;120:384–94. http://dx. doi.org/10.1016/j.supflu.2016.05.044. [22] Jung SJ, Kim SH, Chung IM. Comparison of lignin, cellulose, and hemicellulose contents for biofuels utilization among 4 types of lignocellulosic crops. Biomass Bioenerg 2015;83:322–7. http://dx.doi.org/10.1016/j. biombioe.2015.10.007. [23] Badu M, Twumasi SK, Boadi NO. Effect of lignocellulosic in wood used as substrate on the quality and yield of mushrooms. Food Nutr Sci 2011;2:780–4. http://dx.doi.org/10.4236/fns.2011.27107. [24] Wagner J, Bransgrove R, Beacham TA, Allen MJ, Meixner K, Drosg B, et al. Coproduction of bio-oil and propylene through the hydrothermal liquefaction of polyhydroxybutyrate producing cyanobacteria. Biores Technol 2016;207:166–74. http://dx.doi.org/10.1016/j.biortech.2016.01.114. [25] Chen Y, Mu R, Yang M, Fang L, Wu Y, Wu K, et al. Catalytic hydrothermal liquefaction for bio-oil production over CNTs supported metal catalysts. Chem Eng Sci 2017;161:299–307. http://dx.doi.org/10.1016/j.ces.2016.12.010. [26] Yang C, Jia L, Chen C, Liu G, Fang W. Bio-oil from hydro-liquefaction of Dunaliella salina over Ni/REHY catalyst. Biores Technol 2011;102:4580–4. http://dx.doi.org/10.1016/j.biortech.2010.12.111. [27] Rönnlund I, Myréen L, Lundqvist K, Ahlbeck J, Westerlund T. Waste to energy by industrially integrated supercritical water gasification - Effects of alkali

[28]

[29]

[30]

[31]

[32]

[33]

[34]

[35]

[36]

[37]

[38]

[39]

251

salts in residual by–products from the pulp and paper industry. Energy 2011;36:2151–63. http://dx.doi.org/10.1016/j.energy.2010.03.027. Watanabe M, Iida T, Inomata H. Decomposition of a long chain saturated fatty acid with some additives in hot compressed water. Energy Convers Manage 2006;47:3344–50. http://dx.doi.org/10.1016/j.enconman.2006.01.009. Lazdovica K, Liepina L, Kampars V. Catalytic pyrolysis of wheat bran for hydrocarbons production in the presence of zeolites and noble-metals by using TGA-FTIR method. Biores Technol 2016;207:126–33. http://dx.doi.org/ 10.1016/j.biortech.2016.01.117. Karago S, Bhaskar T, Muto A, Sakata Y. Hydrothermal upgrading of biomass: effect of K2CO3 concentration and biomass/water ratio on products distribution. Biores Technol 2006;97:90–8. http://dx.doi.org/10.1016/j. biortech.2005.02.051. Malins K, Kampars V, Brinks J, Neibolte I, Murnieks R, Kampare R. Bio-oil from thermo-chemical hydro-liquefaction of wet sewage sludge. Biores Technol 2015;187:23–9. http://dx.doi.org/10.1016/j.biortech.2015.03.093. Wildschut J, Melián-Cabrera I, Heeres HJ. Catalyst studies on the hydrotreatment of fast pyrolysis oil. Appl Catal B-Environ 2010;99:298–306. http://dx.doi.org/10.1016/j.apcatb.2010.06.036. Zhang X, Wang T, Ma L, Zhang Q, Jiang T. Hydrotreatment of bio-oil over Nibased catalyst. Biores Technol 2013;127:306–11. http://dx.doi.org/10.1016/j. biortech.2012.07.119. Tanneru SK, Steele PH. Direct hydrocracking of oxidized bio-oil to hydrocarbons. Fuel 2015;154:268–74. http://dx.doi.org/10.1016/ j.fuel.2015.03.080. Karnjanakoma S, Bayua A, Haoc X, Kongparakuld S, Samartd C, Abudulaa A, et al. Selectively catalytic upgrading of bio-oil to aromatic hydrocarbons over Zn, Ce or Ni-doped mesoporous rod-like alumina catalysts. J Mol Catal A-Chem 2016;421:235–44. http://dx.doi.org/10.1016/j.molcata.2016.06.001. Koike N, Hosokai S, Takagaki A, Nishimura S, Kikuchi R, Ebitani K, et al. Upgrading of pyrolysis bio-oil using nickel phosphide catalysts. J Catal 2016;333:115–26. http://dx.doi.org/10.1016/j.jcat.2015.10.022. Remón J, Arauzo J, García L, Arcelus-Arrillaga P, Millan M, Suelves I, et al. Biooil upgrading in supercritical water using Ni-Co catalysts supported on carbon nano fibres. Fuel Process Technol 2016;154:178–87. http://dx.doi.org/ 10.1016/j.fuproc.2016.08.030. Remón J, Arcelus-Arrillaga P, García L, Arauzo J. Production of gaseous and liquid bio-fuels from the upgrading of lignocellulosic bio-oil in sub- and supercritical water: effect of operating conditions on the process. Energy Convers Manage 2016;119:14–36. http://dx.doi.org/10.1016/j. enconman.2016.04.010. Ying X, Limin Z, Jiamin C, Xinghua Z, Longlong M, Tiejun W, et al. One step hydrogenation-esterification of model compounds and bio-oil to alcohols and esters over Raney Ni catalysts. Energy Convers Manage 2016;108:78–84. http://dx.doi.org/10.1016/j.enconman.2015.10.062.